Panel Extension Satellite (PETSAT) - A Novel SatelliteConcept Consisting of Modular, Functional and Plug-in

Panels

Shinichi Nakasuka1, Yoshiki Sugawara1, Hironori Sahara1, Kanichi Koyama2, Takanori Okada2 andChisato Kobayashi2

1 Department of Aeronautics and Astronautics, University of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo, JAPAN,{nakasuka,sugawara,sahara}@space.t.u-tokyo.ac.jp

2 SOHLA,Creation Core 2102, Aramoto-kita 50-5, Higashi-Osaka, JAPAN, {koyama,okada,chisato}@sohla.com

AbstractA novel concept of satellite design, named PETSAT, is proposed in this paper. In this concept, a satellite is made of severalFunctional Panels such as the CPU panel, Battery panel, Communication panel, Attitude control panel, or Thruster panel,each of which has a special dedicated function. By connecting these panels with a reliable connection mechanism in aplug-in fashion, the total integrated system has a full satellite’s function. Various combinations of functional panels, (forexample, two CPU panels + one communication panels + three attitude control panels + two battery panels, etc.) provideflexibility to deal with various mission requirements, even though the basic panels are the same for various missions. Theconcept, technical issues and conceptual study results of PETSAT will be discussed.

1. Concept of PETSATA novel concept of satellite design, named PETSAT, isproposed in this paper. In this concept, a satellite ismade of several Functional Panels each of which has aspecial dedicated function. By connecting these panelsby reliable connection mechanism in plug-in fashion,the total integrated system has a full satellite’s function.Various combinations of different kinds of functionalpanels, sometimes with multiple instances of identicalpanels, provide flexibility to deal with various missionrequirements, even though the basic panels are the samefor various missions. These panels are stowed duringlaunch into a small volume (left figure in Fig. 1), andare extended on orbit (right figure in Fig. 1), potentiallyrealizing a satellite with a large antenna, large solar cellarea or large boom.PETSAT intends to change the satellite development cyclein the following way:

1. Each functional panel can be produced in massquantities so that the reliability can be improved,and the produced panels can be stocked.

2. When a certain mission is given, the satellitebus suitable to the mission requirements can beconfigured only by connecting the appropriatepanels in appropriate quantities in plug-in fashionwithout much effort needed in ground test of thetotal system.

This semi-customizable satellite production process isexpected to dramatically reduce required manufacture andtest time, as well as workload, resulting in a drasticreduction of the individual satellite cost and developmenttime. Mass production of the panels is the key to reducingcost and improving reliability, but in the conventionalsatellite concept, mass production is difficult even atsubsystem level due to the wide variety of missionrequirements. PETSAT tries to make this possible bymodularizing the various basic subsystem functions intostandard panels, and by dealing with the variety of missionrequirements with the quantity of the utilized panels ofdifferent functions. The concept of PETSAT was initiallyproposed at the Satellite Design Contest in 2002, wherethe concept was the winner of the Best Design Award.Five year research project of PETSAT just started in

86

Figure 1: Concept of PETSAT before deployment (left) and after (right)

Figure 2: Various shapes realisable by PETSAT

2003, supported by NEDO subsidy, in collaboration ofUniversity of Tokyo, Osaka University and manufacturers’community of Higashi Osaka. In addition to the basicconcept development and panel and interface designs,we are pursuing the application of the PETSAT conceptto nano-scale thunder observation and remote sensingsatellites In the following sections, first, some technicalissues to be solved to realize PETSAT will be described,followed by the results of conceptual study. In particular,the requirements on the panel internal structure and panelinterfaces will be discussed in detail, and the first designresults to satisfy the requirements will be given.

2. Technical Issues

2.1 Overall Technical Issues

The key technical issues to realize PETSAT include:

1. The question of how to modularize the satellitebus functions into different panels: for example,whether battery and solar panels should beimplemented in a separate functional panel such as

an electric panel or should be equipped in everypanel as standard basic function, etc.

2. Design of a standard panel structure which providesa suitable thermal and structural environment todifferent kinds of functional panels since for massproduction, different types of panels should use thesame structure.

3. Interface between panels: four kinds of interface,mechanical, electrical, information, and thermalinterfaces should be carefully designed so thatthe desired plug-in simplicity of PETSAT can beachieved.

2.2 Requirements on Panel InterfaceAs for the interfaces, the following requirements should besatisfied.

1. Mechanical InterfaceA panel connection and deployment/latchmechanism, which is very reliable, fault tolerant,soft and highly accurate is required. Also it isrequired that the panels can be easily plugged-in.

87

Figure 3: Remote sensing PETSAT

Figure 4: Panel main structure and link module

The accuracy of the angles between panels afterbeing latched should be of a sufficient level toallow the initially planned satellite shape to beachieved. Also the latch mechanism should be soft,which means that some mechanism is required tobrake the panels’ deployment just before they arelatched. Finally, in order to achieve various satelliteshapes (see Fig. 2), the sequence of the paneldeployment should be carefully designed, otherwisethe deployment becomes stacked. Some mechanismis required which assures that, for example, thedeployment of panel A and B can be made afterpanel C is deployed from panel D.

2. Electrical (power) InterfaceIn principle, the electric power required ineach panel should be generated by solar cellsof the same panel, but in many cases powershould be transferred from one panel to anotherpanel. PETSAT should have the capability toautonomously transfer power in relation to thepower generation and consumption in each panel.Reliability of power lines is another important issue,which should be realized by carefully designedredundancy.

3. Information InterfaceCommunication between panels is critical forPETSAT. The information line should be veryreliable and should have enough communicationcapacity to deal with the flow of house keepingdata as well as mission data. Each panel shouldhave a microprocessor able to control both thisinformation traffic and manage the information flowinside the panel. So, the total system becomesa multi-processor system where the architectureto manage such large number of CPUs shouldbe carefully designed so that the strengths ofa distributed system, such as fault tolerance orcapability of grid-computing, can be pursued asmuch as possible. The information line can be eithera wired bus line or an RF line.

4. Thermal InterfaceThermal coupling within and between panels shouldbe made very strong so that the temperaturedifference between each part of satellite is as smallas possible. [1] indicates that this is the bestgeneral strategy for thermal control because (1) thethermal environment is different from mission tomission, and (2) there is not much design freedom

88

Figure 5: Design of panel main structure

Figure 6: OBC panel inside design

for PETSAT thermal control because the surfacesof panels are almost completely covered with solarcells.

3. Example Missions

Before going into technical details, Let usbriefly give descriptions of two example missionsrealizable by PETSAT.

3.1 Remote Sensing

One potential application is remote sensing. Forobtaining high resolution images, the optical systemshould have long focal length, which requiresseveral mirrors for reflections or a certain distancebetween a lens and the imaging device such as aCMOS or CCD. Usually several mirrors are utilizedfor this objective, but high accuracy is requiredfor mirror surfaces, which results in extremelyhigh cost, often with regard to required structuralstrength. The latter requirement is tough formicro/nano satellites because of size limitations.PETSAT can solve the latter problem through itspanel extension mechanism, such as in Fig. 3.Extremely high resolution, such as less than 10 m,

is difficult, but a certain level of resolution suchas 20 m - 50 m is expected to be achievable withvery low cost. Three axis stabilization is required,which is achieved by three attitude control panels,each dealing with stabilization for one axis. In orderto obtain well focused images, the accuracy of theangle between panels after deployment is essentiallyimportant. However this is very hard to achievebecause of the back-rush of the latch or distortion ofthe panel structure due to temperature change, etc.One method to solve this problem is to control theposition and attitude of the image receptor such asCCD or CMOS using very precise actuator. Thismechanism is now under development.

3.2 Interferometric Positioning System

If an RF wave arrives at antennae located at differentpositions on PETSAT, then the difference in arrivaltime, or phase difference, provides information asto the direction of the RF signal source from thePETSAT. If there are three antennae which arenot located collinearly, then the direction of thesignal source in 3D space can be estimated. Theaccuracy of the direction measurement depends onthe distance between the antennae (base line) andthe knowledge of the relative positions of these

89

Figure 7: Severest thermal condition for 2 panels

Figure 8: Heat lane plate

antennae with respect to the satellite body frame.Long base line can be achieved in PETSAT byextension of several panels between the antennaepanels, and the relative positions of antennae can beestimated by calibrating this sensor system using asignal generated from a known point on PETSAT.

If we want to obtain the location of the RF source onthe Earth, then the PETSAT should know its attitudeprecisely. Several types of navigation sensors forattitude can be employed, including IRU (gyros),magnetometers, Earth sensors, sun sensors andstar sensors. However the interferometric sensoritself can be also a precise navigation sensor, i.e.attitude can be estimated by obtaining an RF signalgenerated by a ground station whose position isknown. These two calibration methods are nowunder study.

4. Distributing Functions intoPanels

In PETSAT, some satellite functions may beimplemented as specialized functions of particularpanels, and other functions may be implemented asstandard functions in all types of panel. Therefore,one of the important research issues is how todistribute various satellite functions into differenttypes of panel. To determine this, the followingrequirements for PETSAT features should beobserved:

(a) Interface simplicity: panels of different typescan be plugged-in in any quantity whilesatisfying the requirements on the four typesof interface described in 2.2.

(b) Functions should be improved by increasingthe number of panels: for example, thecommunication capacity should be reinforced

90

Figure 9: Thermal design inside the main panel structure

Figure 10: Results of thermal simulator for two connecting panels as modeled in Fig. 7

by the number of communication panelsemployed

(c) Standard panel structure: the structures ofdifferent panels should be almost the same sothat mass production of the panel structure ispossible.

(d) Flat panels: in order to be deployable, thepanels should be flat, which requires that,for example, only one axis wheel can beimplemented in one panel.

(e) Fault tolerance and graceful degradation: thesatellite functions can be maintained at adegraded level in case of failures of certainpanels or interface components (such asinformation lines, power lines, hinges, etc).

After the examinations taking into account howseveral example missions can be achieved byPETSAT, the following distribution strategy hasbeen found appropriate.

(a) The variety of panels are as follows: ( ) show

the specialized functions implemented in eachpanel

i. OBC Panel (high performance CPU andlarge memory)

ii. Communication Panel (transmitter,receiver and antenna)

iii. Attitude Control Panel (reaction/momentumwheel/magnetic torquer/gyro: one panelfor one axis control)

iv. Orbit Control Panel (thruster, valve, pipesand propellant tank)

v. Battery Panel (large battery)vi. Mission Panel (different for different

missions)

(b) The following components are implemented asstandard functions in all types of panel.

i. solar cell and small batteryii. inter-panel communication-related

components

91

iii. a local CPU and memory which managesinformation inside the panel

5. As to the electrical power, each panel’s solar cellsupplies power to the equipment in the panel, andthe residual power can be transported to any panelthrough negotiation between the CPUs of eachpanel, considering the power demand of each panel

6. With respect to the information management, thereare many CPUs, which should be efficiently utilizedfor realizing fault tolerance, graceful degradation,and grid computing to help in informationprocessing for the mission panel

5. Example Designs of Panels andInterfacesThe PETSAT project is now in the initial conceptual studyphase, during which the following first step design hasbeen obtained.

5.1 Panel StructureThe panel structure should endure the launch environment.PETSAT is more tolerant against acceleration loadbecause panels are stowed flat during launch. Thevibration environment should, however, be consideredseriously because the hinges between panels makethe structural frequency lower. Another importantrequirement is that the same standard structure shouldbe used for different types of panels. Considering therequired mechanical interface described in 2.2, the mainstructure and the hinge-part, named link module should beseparated as in Fig. 4. The link modules are of severaltypes which can latch the panels at different angles (suchas 90, 180, 270 ◦)Fig. 5 shows one example design of the main structure,which consists of several plates and stiffener (placerbetween plates). The central plate, termed the motherboard, has many holes aligned in a regular grid forattaching components to the panels. Fig. 6 shows oneexample of the interior design of a panel (for the OBCpanel). The Bus Controller and battery are inserted as thestandard functions, and CPU and large memory are theSpecialized functions. Solar cells are not shown in thisfigure. In order to improve the heat conductivity withinand between the plates, thermal conductors (carbon sheetand Heat Lane Plate) are inserted. Currently the size ofthe panel is designed as 300 mm by 300 mm with 50 mmheight.

5.2 Thermal DesignAccording to the thermal interface requirements, thetemperature difference between panels has been targetedto be less than 5◦C. As the solar heat input is usually much

larger than the power generated by internal components,the most severe situation, probably yielding the maximumtemperature difference between panels, is the case wheresolar heat is incident on one panel (say, panel P), and theother panel (panel Q), which is 90 ◦ bent from panel Pdoesn’t get any solar heat (Fig. 7). In this case, the300 mm square panel P gets 135 W of solar heat, whichshould be transported to the other side of panel P and theneighboring panel Q.Several heat conducting devices have been examined,and the two candidates, graphite sheet (heat conductivityis about 800W/m2K) and small heat pipe named HeatLane Plate (Fig. 8) are found to be appropriate for heattransfer within and between panels. The Heat Lane Plateis rigid structure and so cannot be used for inter-panelheat transfer. But, it has high heat transport capability(3 × 106 W/m2). If only the graphite sheet is used forthe inter-panel heat transfer, the intra-panel temperaturedifference is so large that it prevents efficient heat transferbetween panels. So the Heat Lane Plate is inserted insidethe panel to lower the intra-panel temperature difference.Fig. 9 shows the example design of inserting a Heat LanePlate and graphite sheet inside the panel.The thermal analysis using a simulator indicated that withthese heat conducting devices, the intra- and inter-paneltemperature difference can be less than 5◦C. Fig. 10shows the thermal simulator results for a simple modelof two panels connected as in Fig. 7. It was shown thatthe temperature difference within one panel is less than5 ◦C and the difference between the two panels at theconnecting point can be suppressed under 5 ◦C.

5.3 Information InfrastructureFirstly, the question of which of the two candidatesfor information lines, wired bus or RF LAN, shouldbe employed has been studied, taking the interfacerequirements into account. Though RF LAN is superiorin that no wiring is required, and so it is more tolerantto panel failure, the possible interaction between RFLAN and PETSAT-ground RF communication and therelatively large power required makes the usage of wiredbus more attractive. With regard to the wired bus system,CAN bus is promising as it is capable of relatively highspeed (1Mbps) and has already been used in many areas,so that there exists much supporting software/hardwarefor development of CAN bus system. Currently aninformation system using CAN bus and an SH-series buscontroller is being developed.

6. ConclusionsA novel satellite concept of PETSAT is proposed, and itsfeatures, applications and basic design philosophy havebeen described. In order to further space developmentactivities, satellite design philosophy needs an essentialbreakthrough so that reliable and capable satellites can be

92

fabricated within a shorter period and at far lower cost.PETSAT can be said to be one trial to approach such abreakthrough.

AcknowledgementsPETSAT research project is supported by NEDO subsidy.

Bibliography[1] Yuya Nakamura, Hideyuki Tanaka, Masaki Nagai, and

Ryu Funase. Multi-objective panel extension satellitepetsat. In 10th Satellite Design Contest, October 2002.